Browse Prior Art Database

Verification of Numeric Handwriting Recognition

IP.com Disclosure Number: IPCOM000075904D
Original Publication Date: 1971-Dec-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 2 page(s) / 55K

Publishing Venue

IBM

Related People

Keillor, RD: AUTHOR [+3]

Abstract

The character recognition errors in an optical character reader can be divided into two categories. 1) A failure to recognize (reject) or a character that is recognized as two or more different characters (con- flict). These types of errors are generally detectable by the optical character reader. 2) An incorrect recognition (substitutions) of a character. This is not generally detectable unless some method of self-checking has been built into the data being scanned. The description and flow chart outline a method to detect, in the optical reader hardware, some of the previously undetected errors. The method, as shown in the flow chart, is based on the premise that character substitutions in Numeric Handwriting, generally have a probability of occurrence of less than 1.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 72% of the total text.

Page 1 of 2

Verification of Numeric Handwriting Recognition

The character recognition errors in an optical character reader can be divided into two categories. 1) A failure to recognize (reject) or a character that is recognized as two or more different characters (con- flict). These types of errors are generally detectable by the optical character reader. 2) An incorrect recognition (substitutions) of a character. This is not generally detectable unless some method of self-checking has been built into the data being scanned. The description and flow chart outline a method to detect, in the optical reader hardware, some of the previously undetected errors. The method, as shown in the flow chart, is based on the premise that character substitutions in Numeric Handwriting, generally have a probability of occurrence of less than 1. That is, a substitution of a particular character will generally not occur each and every time that particular character is read. If then, each character is read more than once and the results of each read are compared, many of the character substitutions will be detectable. The implementation can be by programming or hardware, with the hardware method having the following advantages:

1. Speed--The hardware method eliminates the curve follow for matrix setup on the second read of the character, as well as the artificial follow and CPU compare time. This amounts to a total of approximately 1.5 ms per character.

2. Programming--The hardware verify meth...